Load PyTDC datasets and generate train, valid, and test splits.
npx skills add https://github.com/JosephWoodall/noosphere --skill pytdc-josephwoodall
Please help me install this Agent Skill. Skill: pytdc Source: https://github.com/JosephWoodall/noosphere/tree/main/.agent/skills/pytdc Command: npx skills add https://github.com/JosephWoodall/noosphere --skill pytdc-josephwoodall
PyTDC provides AI-ready datasets and benchmarks for drug discovery and development, offering curated ADME, toxicity, and drug-target datasets with standardized splits and evaluation metrics to accelerate model benchmarking and reproducibility.
Install PyTDC and load a dataset such as ADME(name='Caco2_Wang') and call get_split to obtain train/valid/test partitions.
High-intent search queries and answers about installing and using this skill.
AI-ready datasets for drug discovery are loaded by installing the framework and calling a data loader like ADME(name='Caco2_Wang'). This provides curated therapeutic benchmarks for ADME, toxicity, and DTI tasks.
Reproducible benchmarking requires standardized dataset splits using the get_split function. This yields train, valid, and test partitions through scaffold, cold, or temporal splitting methods for model evaluation.
Molecule generation models can be benchmarked using built-in generation tasks like MolGen, RetroSyn, and PairMolGen. The framework provides generation oracles and utility scripts for end-to-end evaluation.
Single-instance prediction tasks like ADME, Toxicity, and HTS are fully supported, alongside multi-instance tasks including DTI, DDI, and PPI. This enables comprehensive evaluation across diverse drug discovery scenarios.
Pandas and numpy are required dependencies to evaluate ADME and toxicity models. They handle data manipulation for the loaded therapeutic datasets and benchmarking utility scripts.
Standardized therapeutic DTI datasets ensure reproducible model evaluation by providing curated data with built-in evaluation routines. Custom data lacks these standardized splits and metrics, making benchmark comparison difficult.
omicverseClean, reshape, normalize and engineer features from CSV/TSV/Parquet inputs using pandas and numpy locally.
KaiserWhoLearnsTrain and evaluate machine learning models across scikit-learn, PyTorch, and TensorFlow.
Imad-OuteAnalyze and modify molecular structures in cheminformatics workflows.
Estrella-231Load PyTDC therapeutic datasets with scaffold and cold splits for ML modeling.
